Tuberculosis (TB) is a leading infectious cause of death in developing countries. Control of TB has been undermined by changing demographic patterns and the HIV epidemic. The World Health Organization has endorsed directly observed therapy, short course (DOTS) as the principal strategy for controlling TB worldwide. Accumulating evidence suggests the DOTS strategy will be insufficient to produce a reduction in TB incidence and will result in continued high rates of TB deaths for at least a decade. Molecular epidemiologic data suggest that up to one-third of TB in both developing and comprehensive TB control strategies may be more efficacious in controlling the disease. The aim of this study is to conduct a randomized, community trial of alternative strategies to control TB in a developing country with high rates of TB and endemic HIV. Two communities in Rio de Janeiro with similar TB epidemiology have been selected for the trial. In one community, DOTS+ will be used, incorporating DOTS for TB and active outreach to evaluate household contacts and routine multiproject proposal, a randomized trial of weekly rifapentine and isoniazid versus twice weekly rifampin and pyrazinamide study will be the community incidence of TB after four years of intervention and the molecular epidemiologic patterns of TB isolates in the two communities. We hypothesize that DOTS+ will result in an initial increase in TB rate as a result of case finding, but after four years will produce lower case rates than DOTS alone. We also hypothesize that DOTS+ will result in less clonal clustering of TB isolates after four years than DOTS alone, with fewer isolates in clusters and smaller cluster sizes. This innovative strategy for controlling TB in a high-incidence setting will provide important information for TB control strategies worldwide.